Paper Info
Title
Statistical Instance-Based Ensemble Pruning for Multi-class Problems
Abstract
Recent research has shown that the provisional count of votes of an ensemble of classifiers can be used to estimate the probability that the final ensemble prediction coincides with the current majority class. For a given instance, querying can be stopped when this probability is above a specified threshold. This instance-based ensemble pruning procedure can be efficiently implemented if these probabilities are pre-computed and stored in a lookup table. However, the size of the table and the cost of computing the probabilities grow very rapidly with the number of classes of the problem. In this article we introduce a number of computational optimizations that can be used to make the construction of the lookup table feasible. As a result, the application of instance-based ensemble pruning is extended to multi-class problems. Experiments in several UCI multi-class problems show that instance-based pruning speeds-up classification by a factor between 2 and 10 without any significant variation in the prediction accuracy of the ensemble.
Year
DOI
Venue
2009
10.1007/978-3-642-04274-4_10
ICANN (1)
Keywords
Field
DocType
final ensemble prediction,instance-based pruning speeds-up classification,lookup table,multi-class problems,computational optimizations,statistical instance-based ensemble pruning,prediction accuracy,instance-based ensemble pruning procedure,instance-based ensemble pruning,current majority class,multi-class problem,uci multi-class problem,neural networks,ensemble learning,neural network
Lookup table,Data mining,Pattern recognition,Computer science,Artificial intelligence,Ensemble prediction,Artificial neural network,Ensemble learning,Machine learning,Pruning
Conference
Volume
ISSN
Citations 
5768
0302-9743
4
PageRank 
References 
Authors
0.41
7
3
Name
Order
Citations
PageRank
Gonzalo Martínez-Muñoz152423.76
Daniel Hernández-Lobato244026.10
Alberto Suárez348722.33